pptList/scripts/sanitize_music_templates.py

from __future__ import annotations

import os
import sys
import time
import zipfile
import xml.etree.ElementTree as ET
from pathlib import Path

from pptx import Presentation
from pptx.enum.shapes import MSO_SHAPE_TYPE


TARGET_DIR = Path("public/ai-courseware-templates")

TEMPLATES = [
    (
        "fairytale-forest.pptx",
        Path(r"H:\音乐教材PPT\人教版（2024）\1年级下册（新）\【2025新教材】人教版音乐一年级下册-《动画音乐会》-课件.pptx"),
    ),
    ("warm-handdrawn.pptx", Path(r"H:\音乐教材PPT\人音版（2024）\3年级上册（新）\音乐小屋.pptx")),
    (
        "music-doodle.pptx",
        Path(r"H:\音乐教材PPT\人音版（2024）\7年级上册（新）\课件《溜冰圆舞曲》新版人音版七年级上册第二单元.pptx"),
    ),
    ("ink-wash.pptx", Path(r"H:\音乐教材PPT\人音版（2024）\8年级上册（新）\《姑苏行》《十面埋伏》.pptx")),
    ("stage-documentary.pptx", Path(r"H:\音乐教材PPT\人音版（2024）\8年级上册（新）\《鼓乐》.pptx")),
    (
        "modern-geometry.pptx",
        Path(r"H:\音乐教材PPT\人教版（2024）\8年级上册（新）\学习项目二 探索旋律结构的规律.pptx"),
    ),
]


def select_slide_indexes(total: int) -> set[int]:
    """保留正文版式，避开通常包含版权页的首尾页。"""
    if total <= 4:
        return set(range(total))
    candidates = [1, round(total * 0.35), round(total * 0.6), total - 2]
    return {min(total - 1, max(0, index)) for index in candidates}


def delete_shape(shape) -> None:
    element = shape._element
    element.getparent().remove(element)


def clear_shape_text(shape, ordinal: list[int]) -> None:
    if shape.shape_type == MSO_SHAPE_TYPE.GROUP:
        for child in list(shape.shapes):
            clear_shape_text(child, ordinal)
        return
    if shape.shape_type in {MSO_SHAPE_TYPE.MEDIA, MSO_SHAPE_TYPE.WEB_VIDEO}:
        delete_shape(shape)
        return
    if not getattr(shape, "has_text_frame", False):
        return

    text_frame = shape.text_frame
    if not text_frame.text.strip():
        return
    text_frame.clear()
    placeholder = "章节标题" if ordinal[0] == 0 else "在此填写教学内容"
    paragraph = text_frame.paragraphs[0]
    paragraph.text = placeholder
    ordinal[0] += 1


def remove_media_relationships(slide) -> None:
    for relation_id, relation in list(slide.part.rels.items()):
        relation_type = relation.reltype.lower()
        if any(token in relation_type for token in ("audio", "video", "media")):
            slide.part.drop_rel(relation_id)


def remove_notes_parts(path: Path) -> None:
    """删除备注、课件标题等不可见信息，避免原课程信息随模板保留。"""
    temp_path = path.with_suffix(".cleaning.pptx")
    with zipfile.ZipFile(path, "r") as source, zipfile.ZipFile(temp_path, "w", zipfile.ZIP_DEFLATED) as target:
        for entry in source.infolist():
            name = entry.filename
            if name.startswith(("ppt/notesSlides/", "ppt/notesMasters/")) or name == "docProps/app.xml":
                continue
            content = source.read(entry)
            if name.endswith(".rels"):
                root = ET.fromstring(content)
                for relation in list(root):
                    relation_type = relation.attrib.get("Type", "")
                    if relation_type.endswith(("/notesSlide", "/notesMaster", "/extended-properties")):
                        root.remove(relation)
                content = ET.tostring(root, encoding="utf-8", xml_declaration=True)
            elif name == "[Content_Types].xml":
                root = ET.fromstring(content)
                for override in list(root):
                    part_name = override.attrib.get("PartName", "")
                    if "notesSlides" in part_name or "notesMasters" in part_name or part_name == "/docProps/app.xml":
                        root.remove(override)
                content = ET.tostring(root, encoding="utf-8", xml_declaration=True)
            elif name.startswith("ppt/slides/slide") and name.endswith(".xml"):
                root = ET.fromstring(content)
                presentation_ns = "{http://schemas.openxmlformats.org/presentationml/2006/main}"
                drawing_ns = "{http://schemas.openxmlformats.org/drawingml/2006/main}"
                media_shape_tags = {
                    f"{drawing_ns}audioFile",
                    f"{drawing_ns}videoFile",
                    "{http://schemas.microsoft.com/office/powerpoint/2010/main}media"
                }
                for parent in root.iter():
                    for child in list(parent):
                        if child.tag != f"{presentation_ns}pic":
                            continue
                        if any(node.tag in media_shape_tags for node in child.iter()):
                            parent.remove(child)
                for text_node in root.iter("{http://schemas.openxmlformats.org/drawingml/2006/main}t"):
                    if text_node.text and text_node.text.strip():
                        text_node.text = "在此填写教学内容"
                for properties in root.iter(f"{presentation_ns}cNvPr"):
                    properties.attrib.pop("descr", None)
                    properties.attrib.pop("title", None)
                content = ET.tostring(root, encoding="utf-8", xml_declaration=True)
            target.writestr(entry, content)
    for attempt in range(5):
        try:
            os.replace(temp_path, path)
            return
        except PermissionError:
            if attempt == 4:
                raise
            time.sleep(1)


def retain_selected_slides(presentation: Presentation, keep_indexes: set[int]) -> None:
    for index in reversed(range(len(presentation.slides))):
        if index in keep_indexes:
            continue
        relation_id = presentation.slides._sldIdLst[index].rId
        presentation.part.drop_rel(relation_id)
        del presentation.slides._sldIdLst[index]


def sanitize(source: Path, output: Path) -> None:
    presentation = Presentation(source)
    retain_selected_slides(presentation, select_slide_indexes(len(presentation.slides)))
    for slide in presentation.slides:
        ordinal = [0]
        for shape in list(slide.shapes):
            clear_shape_text(shape, ordinal)
        remove_media_relationships(slide)
    presentation.core_properties.author = "系统模板"
    presentation.core_properties.last_modified_by = "系统模板"
    presentation.core_properties.subject = "音乐课件模板"
    presentation.core_properties.keywords = ""
    presentation.core_properties.comments = ""
    presentation.save(output)
    remove_notes_parts(output)


def main() -> None:
    TARGET_DIR.mkdir(parents=True, exist_ok=True)
    if "--clean-output" in sys.argv:
        for output in sorted(TARGET_DIR.glob("*.pptx")):
            remove_notes_parts(output)
            print(f"{output.name}\t{output.stat().st_size / 1024 / 1024:.1f} MB")
        return
    for filename, source in TEMPLATES:
        output = TARGET_DIR / filename
        sanitize(source, output)
        print(f"{filename}\t{output.stat().st_size / 1024 / 1024:.1f} MB")


if __name__ == "__main__":
    main()